Of all the decisions in an IoT program, connectivity is the one that lives longest and is hardest to undo. A device installed in a meter cabinet, an elevator shaft or a farm field may stay in service for ten to fifteen years. If its radio stops being supported halfway through that life, the cost is not a software update but a truck roll to every unit. Many organizations are learning this the hard way as 2G and 3G networks close, while a new generation of options, from 5G RedCap to LoRaWAN with satellite backhaul, promises better economics.
This guide explains what is changing in 2026, how the main technologies compare, and a structured way to choose connectivity that will last as long as your devices do.
The Forces Reshaping IoT Connectivity
Legacy network sunsets are accelerating
According to the Global mobile Suppliers Association (GSA), as reported by Light Reading in August 2026, there were 313 completed, planned or in-progress 2G and 3G switch-offs across 89 countries and territories at the end of June 2026, and Europe accounted for 43 percent of them. Operators want the spectrum and energy savings; IoT users want time. The tension is visible in France, where Orange postponed its plans and now aims to switch off 2G by 2028 and 3G by 2030, partly because of concerns about elevators and other IoT systems still relying on older networks.
For anyone with 2G or 3G devices in the field, the question is no longer whether to migrate, but how quickly and to what.
5G finally has an IoT tier for mid-range devices
5G’s early releases focused on broadband and ultra-reliable low-latency use cases, leaving a gap for mid-range IoT devices that need more than LTE-M but far less than a smartphone. Reduced Capability (RedCap), introduced in 3GPP Release 17, fills that gap with simplified 5G devices: up to 20 MHz bandwidth, single-antenna configurations and no carrier aggregation, with peak rates cited by the GSA of up to 226 Mbps downlink and 120 Mbps uplink. Enhanced RedCap (eRedCap) in Release 18 goes further, capping peak rates at 10 Mbps to reduce cost and power, with longer extended discontinuous reception cycles for battery life.
The positioning matters: RedCap is expected to replace LTE Category 3 and 4 in mid-tier IoT, and eRedCap to replace LTE Cat-1 and Cat-1bis from 2026 onward. ABI Research, as reported by SDxCentral, forecasts around 80 million RedCap module shipments over five years, with eRedCap making up 71 percent of the RedCap module market through 2029.
LPWAN is maturing toward plug-and-play
Low-power wide-area networks remain the workhorse for battery-powered sensors. The LoRa Alliance published a three-year roadmap in June 2026 focused on easier deployment: device migration between networks and capability discovery in 2026; zero-touch onboarding and standardized gateway and network-server interfaces in 2027; and application integration such as a LoRaWAN mapping to OPC UA for industrial use. Improved satellite discovery is also on the 2026 list, reflecting growing interest in hybrid terrestrial and satellite coverage.
The Main Options Compared
No technology wins everywhere. The right choice depends on data rate, power budget, mobility, coverage, and who you want to depend on for the network.
5G RedCap
- Best for: video surveillance, industrial routers and gateways, wearables with rich data, point-of-sale and fixed wireless devices that need tens of megabits or more.
- Strengths: native to 5G standalone networks, long technology runway, support for private 5G and network slicing where operators offer it.
- Watch-outs: depends on 5G standalone coverage, which varies by country and operator; module prices are still higher than mature LTE categories.
eRedCap
- Best for: the large middle of IoT that today uses LTE Cat-1 or Cat-1bis: alarm panels, telematics, connected appliances, smart grid equipment.
- Strengths: 5G longevity at lower complexity and power than RedCap.
- Watch-outs: early in its commercial cycle; verify module availability, certification and operator support in each target market before committing a product line.
LTE-M and NB-IoT
- Best for: battery-powered devices sending small amounts of data: meters, trackers, environmental sensors. LTE-M supports mobility and voice-grade latency; NB-IoT optimizes for deep indoor coverage and very low data rates.
- Strengths: mature modules, power-saving modes, operator-managed networks and roaming agreements in many regions.
- Watch-outs: network availability differs by country and operator (some deploy only one of the two), and roaming support for power-saving features is uneven. Check long-term operator commitments in writing.
LoRaWAN
- Best for: dense sensor deployments on campuses, buildings, farms, utilities and cities, especially where you can own or share gateways.
- Strengths: very low power, unlicensed spectrum, the option of private, public or community networks, and a growing ecosystem of certified devices.
- Watch-outs: low data rates and duty-cycle limits; you may need to operate gateways and a network server yourself; unlicensed spectrum can be congested in busy areas.
Satellite and hybrid
Non-terrestrial network options for NB-IoT and LoRaWAN, as well as proprietary satellite IoT services, are making remote assets reachable without building infrastructure. They usually carry higher per-message costs and latency, so they work best as a fallback for assets that roam out of terrestrial coverage or for very remote sites.
Choose connectivity for the device’s whole life, not for the pilot. A radio that is cheapest today but sunset in year six can make the entire business case negative.
A Six-Step Decision Framework
Use the following sequence to move from requirements to a defensible choice. It works for new products and for migration of legacy fleets.
- Profile the data. Estimate message size, frequency, burstiness and downlink needs (firmware updates count). Note whether video or audio is involved.
- Set the power budget. Mains-powered, rechargeable or primary battery? For batteries, define target life and model energy per message including network attach and update events.
- Map deployment geography and environment. List countries, indoor depth, mobility and whether sites are under your control (where private networks are possible).
- Define the device lifetime and support horizon. Compare it with operator roadmaps and technology sunsets; demand written commitments where the horizon is long.
- Model total cost of ownership. Include modules, certification, data plans or gateway infrastructure, installation, operations and the cost of a future field swap.
- Plan for change. Prefer hardware that supports multiple radio access technologies or modular radios, remote SIM provisioning (eSIM or iSIM) and over-the-air firmware updates, so the device can adapt as networks change.
Quick decision heuristics
- Video, rich telemetry or gateway backhaul on 5G-capable networks: start with RedCap.
- Cat-1 class devices designed for a long life: evaluate eRedCap now, with LTE Cat-1bis as a bridge where eRedCap is not yet available.
- Battery sensors on public networks with mobility: LTE-M; deep indoor, static, very low data: NB-IoT.
- Dense sensors on sites you control: LoRaWAN with private or shared gateways.
- Remote, sparse or roaming assets: terrestrial first, satellite as fallback.
Contracts, Roaming and Private Networks
Technology comparisons get most of the attention, but many connectivity failures are commercial rather than technical. A device can be perfectly engineered and still be stranded by a contract that expires, a roaming agreement that excludes power-saving features, or an operator that quietly stops supporting a band. Treat the connectivity contract as part of the product design.
What to negotiate with connectivity providers
- Technology support commitments: written minimum support dates for each radio access technology and band your devices use, with notice periods for any change long enough to plan field work.
- Roaming scope: confirmation that LTE-M or NB-IoT roaming, including power-saving modes, is available in every country you deploy, rather than only basic data roaming.
- Remote SIM provisioning: support for eSIM or iSIM profile switching under the current GSMA IoT specifications, so you can change provider without touching devices.
- Data and visibility: APIs for SIM status, usage and network events, which feed your device management platform and help spot failing units early.
- Exit terms: how quickly and at what cost you can move profiles to another provider if service or pricing deteriorates.
When a private network makes sense
Private LTE or 5G networks, and privately operated LoRaWAN, give you control over coverage, capacity and lifecycle on sites you own: factories, ports, mines, campuses and large logistics centers. They make most sense when you need guaranteed coverage in difficult radio environments, deterministic performance for many devices, or data that must stay on site. The trade-offs are spectrum access (which differs by country), the skills to run a radio network, and the need to plan for devices that also leave the site. Many organizations end up with a hybrid: a private network on site, with public networks and roaming SIMs for everything that moves.
Field testing before you commit
Before freezing a design, run a structured field test. Install candidate devices in the worst realistic locations rather than the best, log signal quality, attach times, message success rates and energy per message over several weeks, and include at least one firmware update over the air. Compare results per operator and per technology. A few weeks of disciplined measurement routinely prevents years of support tickets.
Migrating a Legacy 2G/3G Fleet
Sunset migrations are logistics projects with an engineering core. A disciplined approach reduces cost and risk:
- Build an accurate inventory. Identify every device by radio technology, operator, location, firmware and customer. Network usage records from your connectivity provider help find devices nobody remembers.
- Map sunset dates. Compile operator switch-off dates per country and prioritize fleets with the earliest deadlines and highest business impact (safety alarms, elevators, payments).
- Choose the replacement path per segment. Some devices can take a plug-in radio upgrade; others need full replacement or a gateway that aggregates local devices onto a new backhaul.
- Pilot and certify. Validate coverage, power and operator certification in each market before ordering at scale.
- Schedule field work smartly. Combine radio swaps with planned maintenance visits and battery replacements to avoid extra truck rolls.
- Decommission cleanly. Retire SIMs, update asset records and confirm that platform-side logic no longer expects legacy devices.
A Worked Example: A Building Services Provider
Consider a generic building services company that maintains elevator emergency phones, fire-panel communicators and energy meters across several European countries. Most communicators use 2G, and operators in two of its markets have announced switch-offs within three years.
The company profiles its devices: emergency phones need reliable voice or voice-equivalent connectivity and are safety critical; fire-panel communicators send small alarms but need high availability; meters send small readings a few times a day. It chooses LTE-M with fallback to another operator via multi-network SIMs for the communicators and emergency phones, prioritizing them by sunset date. For meters inside the buildings it already services, it deploys LoRaWAN gateways on site, connected over the building’s internet link with LTE backup, and replaces meter radios during scheduled meter exchanges.
For a new generation of gateways planned for the following year, the company evaluates eRedCap modules in its two largest markets and keeps LTE Cat-1bis as a fallback where eRedCap support is not yet commercial. Every new device ships with remote SIM provisioning and over-the-air firmware updates. The program turns a forced migration into an opportunity to standardize hardware and reduce long-term connectivity costs.
Pitfalls to Avoid
- Trusting coverage maps alone. Test in real installation locations, especially basements, shafts and metal enclosures.
- Ignoring power-saving behavior on roaming. Features that make batteries last on a home network may not be honored abroad.
- Underestimating certification. Operator and regulatory certification can add months; plan it into product timelines.
- Forgetting security and regulation. Radios, SIM management and firmware updates are part of the product’s security posture, which the EU Cyber Resilience Act now makes a legal obligation for products sold in the EU.
- Single-network dependency. Without multi-operator or multi-technology options, one operator’s decision can strand your fleet.
Outlook
Over the next few years, expect 5G standalone coverage to broaden, eRedCap modules to become mainstream for mid-tier devices, LoRaWAN deployments to get easier through standardized onboarding and roaming between networks, and satellite to become a routine fallback rather than an exotic option. Ambient, batteryless IoT devices standardized in 3GPP Release 19 will add another layer for very low-cost tagging. The organizations best positioned will be those with flexible hardware, remote provisioning and a clear inventory of what is deployed where.
Regulation will also shape connectivity choices. Security obligations for connected products, spectrum policy decisions about refarming legacy bands, and public-sector requirements for critical infrastructure all influence which technologies operators keep investing in. Building a small internal capability, or a trusted partner relationship, that tracks these signals and translates them into product roadmap decisions is far cheaper than discovering a sunset notice after devices have shipped.
Conclusion
IoT connectivity is a lifetime decision. With 2G and 3G closing and RedCap, eRedCap and LPWAN maturing, 2026 is a good moment to reset your strategy: profile data and power needs honestly, align device lifetimes with network roadmaps, model total cost including future field work, and design for change. Doing so protects both your fleet and your business case.
How Hanoi Technology Can Help
Hanoi Technology helps organizations assess connectivity options, plan 2G/3G sunset migrations and build device and platform software that adapts as networks change.